Our Creativity
On occasion, innovative approaches are needed for the optimal design or analysis of a project. We do this only when existing methods are inadequate for a specfic project. We have neither the interest nor the inclination to develop new methods solely to develop new methods.
We have extensive experience applying existing methods to new applications, extending existing methods, and developing new statistical methods when essential to address real world problems. This work has appeared in the peer-reviewed scientific literature. Examples include:
- applying the Gail-Simon test when the overall results are not statistically significant, arising from work on a 2x2 factorial study (Stat Med 2010;29:210-218) Preprint here
- extending existing stopping rules for studies involving more than two treatment groups for a study involving two active treatments against a single control group, arising from work on a study with this design (Eval Health Prof 2007;30:284-299)
- extending an existing approach to assess whether binary decision rules work better than chance for the analysis of a single group clinical study (Biometrical J 1999;41:25-31)
- developing a new method to incorporate population overlap from consecutive cohorts in the analysis of HEDIS report card data (Jt Comm J Qual Improv 1997;23:299-311; Jt Comm J Qual Improv 1999;25:641-650)
- developing a new method to incorporate assay sensitivity in estimates from population surveys for genetic screening, arising from a genetic screening project (Am J Med Genet 1994;49:317-322; Biometrical J 1996;38:375-384)
- developing a new method to analyze virus neutralization assays incorporating results from the virus challenge dose, arising from work determining the risk posed from a new virus strain (Stat Med 1992;11:1253-1262)
- applying mixture models to determine a criterion for seropositivity in a laboratory assay (J Virol Methods 1990;27:135-144)
- applying mixture models to screen multiple statistical tests for time trends in national mortality data (Stat Med 1988;7:1031-1043)
- developing a new method to determine sample size for matched case-control studies, arising from work on a national case-control study (Biometrics 1986;42:919-926)
A list of some of our peer-reviewed statistical articles is here and a brief resume for Dr. Parker is here.
For additional information, please contact us.